Virtual power plant energy storage configuration method considering source load uncertainty

Abstract

The invention provides a virtual power plant energy storage configuration method considering source load uncertainty, and the method comprehensively considers the minimum daily operation cost of a virtual power plant as a target, and is a virtual power plant energy storage configuration method based on two-stage robust optimization. According to the method, a two-stage problem is converted into a main problem and a sub-problem in a mixed integer linear form to be solved, and uncertainty adjustment parameters are introduced, so that the energy storage capacity of the virtual power plant can be configured through different conservative degrees in construction, and the energy storage capacity of the virtual power plant can be configured under the background that the permeability of renewable energy sources is continuously improved. The new energy consumption rate is improved, the energy structure is optimized, and the development of a novel power system is promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy management, and in particular relates to a virtual power plant energy storage configuration method based on source-load uncertainty. Background Art

[0003] As a key technology carrier and effective means to improve power demand management, promote new energy consumption, and enhance system regulation flexibility, virtual power plants can effectively promote the development of source-grid-load-storage integration and promote the real-time balance of the power system. Therefore, studying the energy storage configuration method of virtual power plants that takes into account source-load uncertainty has important engineering application value. Summary of the invention

[0004] In view of the shortcomings in the prior art, the present invention provides a virtual power plant energy storage configuration method taking into account the uncertainty of source and load, comprehensively considers the goal of minimizing the comprehensive daily operating cost of the virtual power plant, and proposes a two-stage robust optimization model, which is conducive to promoting the green and low-carbon development of the power industry and optimizing the energy structure.

[0005] The present invention achieves the above technical objectives through the following technical means.

[0006] A virtual power plant energy storage configuration method based on source-load uncertainty comprises the following steps:

[0007] Step 1: Construct a virtual power plant energy storage capacity configuration model. The model aims to minimize the comprehensive daily operating cost of the virtual power plant. Constraints include energy storage power station constraints, photovoltaic output constraints, and industrial and commercial user load constraints.

[0008] Step 2: Based on the deterministic optimization model constructed in step 1, introduce uncertainty factors and convert it into a two-stage robust optimization model with uncertainty factors. The first stage problem of the two-stage model is the outer min problem, and the second stage model is the inner max-min problem. The goal of the model is to find the worst uncertainty scenario and calculate the most economically efficient energy storage capacity configuration plan under the worst scenario;

[0009] Step 3: Based on the duality principle, the two-stage max-min problem of the two-stage robust optimization model established in step 2 is converted into a dual problem, so that the original two-stage max-min problem becomes a max-max problem and is merged into a single max problem, thereby simplifying the entire model and making it easier to handle and solve;

[0010] Step 4: Use the column constraint generation algorithm to solve the two-stage distributed robust optimization model to obtain the constructed virtual power plant energy storage configuration result.

[0011] Furthermore, in step 1, the objective function of the energy storage capacity configuration model of the virtual power plant is:

[0012] min(C1+C2-C3+C4-C5+C6)(1)

[0013] Among them, C1 is the energy storage construction cost; C2 is the energy storage operation and maintenance cost; C3 is the recovery value of energy storage; C4 is the cost of industrial and commercial users of the virtual power plant purchasing electricity from energy storage; C5 is the income of industrial and commercial users of the virtual power plant selling electricity from their own photovoltaic power plants to energy storage power stations; C6 is the cost of industrial and commercial users of the virtual power plant purchasing electricity from the power grid.

[0014] The expression of the construction cost of virtual power plant energy storage is:

[0015] C1=c p P max +c e E max (2)

[0016] In the formula, c p and c e Respectively represent the unit power cost and unit capacity cost of energy storage equipment; P max Indicates the rated power of energy storage; E max It represents the rated capacity of the energy storage.

[0017] The operation and maintenance costs of the energy storage system mainly include labor costs, maintenance expenses, and the replacement costs of some components of the energy storage unit. The expression is:

[0018]

[0019] Where c0 is the annual operation and maintenance cost per unit charge / discharge power of the energy storage system; Y is the service life of the energy storage system; t r is the inflation rate; d r is the discount rate.

[0020] The recycling and reuse value of the virtual power plant energy storage device is directly related to its initial purchase cost, and its expression is:

[0021] C3=γC1 (4)

[0022] Where γ is the recycling value coefficient.

[0023] The comprehensive daily cost of virtual power plant energy storage should also include: the cost of industrial and commercial users of the virtual power plant purchasing electricity from energy storage, the income of self-built photovoltaic power plants of industrial and commercial users of the virtual power plant selling electricity to energy storage power stations, and the cost of industrial and commercial users of the virtual power plant purchasing electricity from the power grid. The expressions are:

[0024]

[0025] in, is the power supplied by the energy storage to industrial and commercial user i, ω1(t) is the price of electricity purchased by the user from the energy storage; is the power supplied by PV j to the energy storage, ω2(t) is the price of electricity purchased by the virtual power plant energy storage from the user's PV; is the power supplied by the grid to industrial and commercial user i, and ω3(t) is the price of electricity purchased by the user from the grid.

[0026] Furthermore, in step 1, the constraints of the energy storage capacity configuration model of the virtual power plant are as follows:

[0027] The constraints related to virtual power plant energy storage are as follows:

[0028] 0≤P ch (t)≤U ch (t)P max (8)

[0029] 0≤P dis (t)≤U dis (t)P max (9)

[0030] U ch (t)+U dis (t)≤1 (10)

[0031]

[0032] P max =0.5E max (13)

[0033] E(0)=0.4E max (14)

[0034]

[0035] Formulas (8)-(9) are the upper and lower limits of the charging and discharging power of energy storage, where P ch (t), P dis (t) is the energy storage charging and discharging power, P max is the maximum charge and discharge power of energy storage; Formula (10) is the charge and discharge state constraint, where U ch (t), U dis (t) is a Boolean variable, which indicates the energy storage charging and discharging state, and is used to ensure that the energy storage cannot be charged and discharged at the same time. The variable is 1, which indicates that the energy storage is charging or discharging, and vice versa. Formula (11) is the capacity conservation constraint of the scheduling period, where η is the energy storage charging and discharging efficiency (the value is 0.95), N Tis the scheduling period (value is 24); Formula (12) is the remaining capacity constraint of energy storage in different time periods, where E(0) is the initial capacity of the energy storage power station, E max is the maximum capacity of the energy storage station; Formula (13) is the energy storage capacity and power configuration constraints; Formula (14) indicates that the daily initial power of the energy storage station is 0.4 times its rated capacity; Formulas (15)-(16) are the balance constraints of the charging and discharging power of the energy storage station, where: The power that the grid supplies to the energy storage power station, is the power supplied by photovoltaic power to energy storage, The power supplied by the energy storage power station to industrial user i.

[0036] The photovoltaic output constraints are as follows:

[0037]

[0038] Formula (17) is the power balance constraint for the photovoltaic output of each industrial and commercial user, that is, photovoltaic power supplies to industrial and commercial users and also sells electricity to energy storage; Formulas (18)-(19) are the upper and lower limit constraints of each power, where P m,max The upper limit of the interconnection line power.

[0039] The load constraints for industrial and commercial users are as follows:

[0040]

[0041] Formula (20) is the power balance constraint of each industrial user load, that is, the user can be powered by its own photovoltaic power, purchase electricity from energy storage and purchase electricity from the grid; Formulas (21)-(22) are the upper and lower limit constraints of each power.

[0042] Furthermore, in step 2, the two-stage robust optimization model containing uncertainty factors is:

[0043] Assuming that the uncertainty factors of photovoltaic output and load power are not considered, a deterministic optimization model for energy storage configuration problem can be constructed, and its compact form is as follows:

[0044]

[0045] In the formula, x and y are optimization variables, and the specific expression is shown in formula (24).

[0046]

[0047] By rewriting equation (1), we can get the coefficient column vector form of the objective function, which is c; A, B, C, D, E, and F are the corresponding coefficient matrices under the equality and inequality constraints respectively; a, b, c, d, e, and f are the constant column vectors corresponding to the constraints respectively. and y are uncertain variables, the details are as follows:

[0048]

[0049] Where P pvA (t),P pvB (t),P pvC (t),P pv1 (t),P pv2 (t) represents the predicted value of photovoltaic output at time t, and P pv,j (t) Replace; P user1 (t),P user2 (t) represents the predicted value of the charging station load power at time t, and P user,Y (t) Replacement.

[0050] Furthermore, in step 2, the two-stage robust optimization model containing uncertainty factors is as follows:

[0051]

[0052]

[0053] In the formula, the first-stage problem is the outer min problem, and the optimized variable is x; the second-stage problem is the inner max-min problem, and the optimized variables are u and y; among them, the minimization problem is to reduce the daily cost of users. The goal of this model is to find the most economically efficient capacity configuration of energy storage when u takes the worst case value (u∈U), that is, to calculate the minimum cost under the worst case. Formula (27) is the uncertainty factor introduced in the original model. By defining the box-type uncertainty set U, the fluctuation range of photovoltaic power generation output and charging station load power can be defined, where P pv,j (t) and P user,Y (t) is the uncertain variable of photovoltaic output and charging station load introduced after considering the uncertainty, and Indicates the fluctuation range of the corresponding prediction value error. Formula (28) indicates that when a set of (x,u) is given, the specific value interval of the optimization variable y is Ω(x,u).

[0054] Furthermore, in step 3, the decoupled main problem of the two-stage robust optimization problem is as follows:

[0055]

[0056] Where k is the current number of iterations; y l is the solution of the subproblem after the lth iteration; u l *It is the value of the uncertain variable u in the worst scenario found after the lth iteration.

[0057] Furthermore, in step 3, the decoupled sub-problems of the two-stage robust optimization problem and its dual problem are as follows:

[0058]

[0059]

[0060] In the formula, Ω(x,u) is the value range of the decision variable y, and γ, λ, ν, π, and β represent the dual variables corresponding to each constraint in the subproblem. Based on the duality principle, the inner min problem of the max-min problem in the subproblem can be converted into an equivalent max problem, thereby simplifying the structure of the model. The dual model of the converted subproblem is as follows:

[0061]

[0062] The uncertain variable corresponding to the optimal solution of the dual problem is a point in the uncertainty set U. When equation (32) takes the maximum value, the value of the uncertain variable u should be located at the boundary of the interval, that is, the worst scenario. Therefore, the uncertainty set defined in the present invention can be transformed into the following:

[0063]

[0064] In the formula, B=[B pv,j (t),B user,Y (t)] T is a binary variable. When the value is 1, the uncertain variable takes the boundary value of the corresponding period in the interval; Γ pv,j With Γ user,Y It is an uncertainty value, indicating the cumulative number of photovoltaic output and charging station load reaching the boundary of the fluctuation range, and its value range is 0~N T All integers within .

[0065] Bringing the transformed uncertainty set into the transformed subproblem, we get the complete dual subproblem, as follows:

[0066]

[0067] In the formula, Indicates the fluctuation range of an uncertain quantity.

[0068] Furthermore, in step 4, the column constraint generation algorithm is used to solve the model as follows:

[0069] Based on the convergence principle of the column constraint generation method, the main problem and subproblems are solved by calling the GUROBI loop iteration on the MATLAB platform. When the upper bound and the lower bound of the original problem meet the convergence principle at the same time, the solution is completed and the energy storage configuration capacity E is obtained. max , P max , the solution process is:

[0070] Step 4.1: Set LB = -∞, UB = +∞, number of iterations k = 1, and convergence accuracy ε ≤ 0.01;

[0071] Step 4.2 Solve the main problem and find the optimal solution Update LB = α;

[0072] Step 4.3 Given the solution Get the objective function value of the subproblem and the value of uncertainty u in the corresponding worst scenario Update upper bound

[0073] Step 4.4: When |UB-LB|≤ε, terminate the iteration and obtain the optimal solution; otherwise, set k=k+1, add the relevant constraints to the main problem, and repeat steps (2) to (4);

[0074] Wherein, the termination condition is |UB-LB|≤ε.

[0075] The present invention has the following beneficial effects:

[0076] The present invention constructs a virtual power plant energy storage configuration model, adopts a two-stage robust optimization principle to deal with the uncertainty of source and load, then simplifies the model by using the strong duality principle, and finally uses the column constraint generation method to solve it on the MATLAB platform to obtain a virtual power plant energy storage configuration scheme that takes into account the uncertainty of source and load, optimizes the energy structure of the virtual power plant, improves the new energy consumption rate of the virtual power plant and the power system, and promotes my country's energy transformation.

[0077] The present invention studies the energy storage configuration method of a virtual power plant under uncertain source and load conditions, achieves the minimum comprehensive daily cost of the virtual power plant, ensures the reliability and economy of the virtual power plant planning scheme, improves the flexibility of the virtual power plant and the system as a whole, optimizes the new energy absorption capacity, and promotes the development of the virtual power plant industry and the construction of new power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 A two-stage solution flow chart based on the column constraint generation algorithm provided by the present invention;

[0079] Figure 2Load demand curves for commercial and industrial users of virtual power plants and electric vehicle charging stations;

[0080] Figure 3 Self-built photovoltaic output prediction curves for virtual power plant commercial and industrial users and electric vehicle charging stations; DETAILED DESCRIPTION

[0081] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0082] The virtual power plant energy storage configuration method taking into account source and load uncertainty of the present invention comprises the following steps:

[0083] Step 1: Build a virtual power plant energy storage capacity configuration model;

[0084] Step 1.1: The energy storage capacity configuration model aims to minimize the comprehensive daily operating cost of the virtual power plant. The constraints include energy storage power station constraints, photovoltaic output constraints, and industrial and commercial user load constraints. The specific construction method is as follows:

[0085] Step 1.1.1: The objective function of the energy storage capacity configuration model is:

[0086] min(C1+C2-C3+C4-C5+C6)(1)

[0087] Among them, C1 is the energy storage construction cost; C2 is the energy storage operation and maintenance cost; C3 is the recovery value of energy storage; C4 is the cost of industrial and commercial users of the virtual power plant purchasing electricity from energy storage; C5 is the income of industrial and commercial users of the virtual power plant selling electricity from their own photovoltaic power plants to energy storage power stations; C6 is the cost of industrial and commercial users of the virtual power plant purchasing electricity from the power grid.

[0088] The expression of the construction cost of virtual power plant energy storage is:

[0089] C1=c p P max +c e E max (2)

[0090] In the formula, c p and c e Respectively represent the unit power cost and unit capacity cost of energy storage equipment; P max Indicates the rated power of energy storage; E max It represents the rated capacity of the energy storage.

[0091] The operation and maintenance costs of the energy storage system mainly include labor costs, maintenance expenses, and the replacement costs of some components of the energy storage unit. The expression is:

[0092]

[0093] Where c0 is the annual operation and maintenance cost per unit charge / discharge power of the energy storage system; Y is the service life of the energy storage system; t r is the inflation rate; d r is the discount rate.

[0094] The recycling and reuse value of the virtual power plant energy storage device is directly related to its initial purchase cost, and its expression is:

[0095] C3=γC1 (4)

[0096] Where γ is the recycling value coefficient.

[0097] The comprehensive daily cost of virtual power plant energy storage should also include: the cost of industrial and commercial users of the virtual power plant purchasing electricity from energy storage, the income of self-built photovoltaic power plants of industrial and commercial users of the virtual power plant selling electricity to energy storage power stations, and the cost of industrial and commercial users of the virtual power plant purchasing electricity from the power grid. The expressions are:

[0098]

[0099] in, is the power supplied by the energy storage to industrial and commercial user i, ω1(t) is the price of electricity purchased by the user from the energy storage; is the power supplied by PV j to the energy storage, ω2(t) is the price of electricity purchased by the virtual power plant energy storage from the user's PV; is the power supplied by the grid to industrial and commercial user i, and ω3(t) is the price of electricity purchased by the user from the grid.

[0100] Step 1.1.2: Determine the constraints of the energy storage capacity configuration model, which include: energy storage power station constraints, photovoltaic output constraints, and industrial and commercial user load constraints, as follows:

[0101] (1) The constraints related to energy storage in virtual power plants are as follows:

[0102] 0≤P ch (t)≤U ch (t)P max (8)

[0103] 0≤P dis (t)≤U dis (t)P max (9)

[0104] U ch (t)+U dis (t)≤1 (10)

[0105]

[0106]

[0107] P max =0.5E max (13)

[0108] E(0)=0.4E max (14)

[0109]

[0110] Formulas (8)-(9) are the upper and lower limits of the charging and discharging power of energy storage, where P ch (t), P dis (t) is the energy storage charging and discharging power, P max is the maximum charge and discharge power of energy storage; Formula (10) is the charge and discharge state constraint, where U ch (t), U dis (t) is a Boolean variable, which indicates the energy storage charging and discharging state, and is used to ensure that the energy storage cannot be charged and discharged at the same time. The variable is 1, which indicates that the energy storage is charging or discharging, and vice versa. Formula (11) is the capacity conservation constraint of the scheduling period, where η is the energy storage charging and discharging efficiency (the value is 0.95), N T is the scheduling period (value is 24); Formula (12) is the remaining capacity constraint of energy storage in different time periods, where E(0) is the initial capacity of the energy storage power station, E max is the maximum capacity of the energy storage station; Formula (13) is the energy storage capacity and power configuration constraints; Formula (14) indicates that the daily initial power of the energy storage station is 0.4 times its rated capacity; Formulas (15)-(16) are the balance constraints of the charging and discharging power of the energy storage station, where: The power that the grid supplies to the energy storage power station, is the power supplied by photovoltaic power to energy storage, The power supplied by the energy storage power station to industrial user i.

[0111] (2) The photovoltaic output constraints are as follows:

[0112]

[0113] Formula (17) is the power balance constraint for the photovoltaic output of each industrial and commercial user, that is, photovoltaics can supply power to industrial and commercial users, and can also sell electricity to energy storage; Formulas (18)-(19) are the upper and lower limit constraints of each power, where P m,max The upper limit of the interconnection line power.

[0114] (3) The load constraints for industrial and commercial users are as follows:

[0115]

[0116] Formula (20) is the power balance constraint of each industrial user load, that is, the user can be powered by its own photovoltaic power, purchase electricity from energy storage and purchase electricity from the grid; Formulas (21)-(22) are the upper and lower limit constraints of each power.

[0117] Step 2: Based on the deterministic optimization model constructed in step 1, introduce uncertainty factors and convert it into a two-stage robust optimization model with uncertainty factors. The first stage problem of the two-stage model is the outer min problem, and the second stage model is the inner max-min problem. The goal of the model is to find the worst uncertainty scenario and calculate the most economically efficient energy storage capacity configuration plan under the worst scenario;

[0118] Step 2.1: Based on the deterministic optimization model constructed in step 1, introduce uncertainty factors and convert it into a two-stage robust optimization model with uncertainty factors;

[0119] Step 2.1.1: Write the deterministic optimization model of the energy storage configuration problem of the plant constructed in step 1 into a compact form as follows:

[0120]

[0121] In the formula, x and y are optimization variables, and the specific expression is shown in formula (24).

[0122]

[0123] By rewriting equation (1), we can get the coefficient column vector form of the objective function, which is c; A, B, C, D, E, and F are the corresponding coefficient matrices under the equality and inequality constraints respectively; a, b, c, d, e, and f are the constant column vectors corresponding to the constraints respectively. and y are uncertain variables, the details are as follows:

[0124]

[0125] Where P pvA (t),P pvB (t),P pvC (t),P pv1 (t),P pv2 (t) represents the predicted value of photovoltaic output at time t, and P pv,j (t) Replace; P user1 (t),P user2 (t) represents the predicted value of the charging station load power at time t, and P user,Y (t) Replacement.

[0126] Step 2.1.2: Introduce uncertainty factors and transform the deterministic model into a two-stage robust optimization model with uncertainty factors:

[0127]

[0128] In the formula, the first-stage problem is the outer min problem, and the optimized variable is x; the second-stage problem is the inner max-min problem, and the optimized variables are u and y; among them, the minimization problem is to reduce the daily cost of users. The goal of this model is to find the most economically efficient capacity configuration of energy storage when u takes the worst case value (u∈U), that is, to calculate the minimum cost under the worst case. Formula (27) is the uncertainty factor introduced in the original model. By defining the box-type uncertainty set U, the fluctuation range of photovoltaic power generation output and charging station load power can be defined, where P pv,j (t) and P user,Y (t) is the uncertain variable of photovoltaic output and charging station load introduced after considering the uncertainty, and Indicates the fluctuation range of the corresponding prediction value error. Formula (28) indicates that when a set of (x,u) is given, the specific value interval of the optimization variable y is Ω(x,u).

[0129] Step 3: Based on the duality principle, the two-stage max-min problem of the two-stage robust optimization model established in step 2 is converted into a dual problem, so that the original two-stage max-min problem becomes a max-max problem and is merged into a single max problem, thereby simplifying the entire model and making it easier to handle and solve;

[0130] Step 3.1: Decouple the two-stage robust optimization model established in step 2 to obtain the main problem (min problem) and sub-problem (max-min problem):

[0131] Step 3.1.1: The decoupled main problem of the two-stage robust optimization problem is as follows:

[0132]

[0133] Where k is the current number of iterations; y l is the solution of the subproblem after the lth iteration; u l * It is the value of the uncertain variable u in the worst scenario found after the lth iteration.

[0134] Step 3.1.2: The decoupled sub-problems and their dual problems of the two-stage robust optimization problem are as follows:

[0135]

[0136] In the formula, Ω(x,u) is the value range of the decision variable y, and γ, λ, ν, π, and β represent the dual variables corresponding to each constraint in the subproblem. Based on the duality principle, the inner min problem of the max-min problem in the subproblem can be converted into an equivalent max problem, thereby simplifying the structure of the model. The dual model of the converted subproblem is as follows:

[0137] Step 3.2: Based on the duality principle, the two-stage max-min problem of the two-stage robust optimization model established in step 2 is converted into a dual problem, thereby simplifying the model:

[0138] Step 3.2.1: The two-stage max-min problem is simplified based on the strong duality principle as follows:

[0139] After simplification, the max-min problem in the subproblem can be converted into an equivalent max problem based on the duality principle, as follows:

[0140]

[0141] The uncertain variable corresponding to the optimal solution of the dual problem is a point in the uncertainty set U. When equation (32) takes the maximum value, the value of the uncertain variable u should be located at the boundary of the interval, that is, the worst scenario. Therefore, the uncertainty set defined in the present invention can be transformed into the following:

[0142]

[0143] In the formula, B=[B pv,j (t),B user,Y (t)] T is a binary variable. When the value is 1, the uncertain variable takes the boundary value of the corresponding period in the interval; Γ pv,j With Γ user,Y It is an uncertainty value, indicating the cumulative number of photovoltaic output and charging station load reaching the boundary of the fluctuation range, and its value range is 0~N T All integers within .

[0144] Bringing the transformed uncertainty set into the transformed subproblem, we get the complete dual subproblem, as follows:

[0145]

[0146] In the formula, Indicates the fluctuation range of an uncertain quantity.

[0147] Step 4: Use the column constraint generation algorithm to solve the two-stage distributed robust optimization model to obtain the constructed virtual power plant energy storage configuration result.

[0148] Step 4.1: Set LB = -∞, UB = +∞, number of iterations k = 1, and convergence accuracy ε ≤ 0.01;

[0149] Step 4.2 Solve the main problem and find the optimal solution Update LB = α;

[0150] Step 4.3 Given the solution Get the objective function value of the subproblem and the value of uncertainty u in the corresponding worst scenario Update upper bound

[0151] Step 4.4: When |UB-LB|≤ε, terminate the iteration and obtain the optimal solution; otherwise, set k=k+1, add the relevant constraints to the main problem, and repeat steps (2) to (4);

[0152] Wherein, the termination condition is |UB-LB|≤ε.

[0153] In order to verify the effectiveness of the present invention, the following implementation example is set:

[0154] Assume that the virtual power plant includes three industrial users, two electric vehicle charging station users and an energy storage system to be configured. The daily load curves of the industrial users and electric vehicle charging stations are as follows: Figure 2 As shown; Each industrial and commercial user and charging station has its own photovoltaic power generation system, and its output prediction curve is as follows Figure 3 As shown; the parameter values ​​related to industrial users and electric vehicle charging stations are shown in Table 1.

[0155] Table 1 Values ​​of relevant parameters for power users

[0156]

[0157] In order to verify the impact of uncertainty on the economic efficiency of the energy storage configuration scheme, this example lists four groups of uncertainty values ​​of photovoltaic and charging stations for analysis and comparison. The results are shown in Table 2:

[0158] Table 2 Virtual power plant and energy storage capacity under different uncertainty values

[0159]

[0160] As can be seen from Table 2, the daily cost and energy storage capacity configuration of the virtual power plant increase with the increase of uncertainty. When the virtual power plant formulates its day-ahead dispatch plan, the number of uncertainties considered is inversely proportional to the risk of the plan and proportional to the cost; the amount of electricity purchased by users from the power grid and energy storage increases with the increase of uncertainty. This is because as the Γ value increases, the power gap between photovoltaics and loads increases, the income from selling electricity decreases, and the amount of electricity purchased and the cost continue to rise.

[0161] In addition to the uncertainty of the value, the error of power prediction will also have a certain impact on the economy. pvX =6 and Γ userY =6 as an example, the impact of different prediction errors on economic performance is shown in Table 3.

[0162] Table 4 Daily cost of virtual power plant and energy storage capacity configuration under different prediction errors

[0163]

[0164] It can be seen that as the prediction error increases, the user's daily cost and the required energy storage capacity also increase.

[0165] At this point, the method proposed by the present invention is implemented.

[0166] According to the theoretically derived formula and simulation results, the following conclusions can be drawn:

[0167] The present invention takes into account the capacity configuration process of virtual power plant energy storage taking into account source and load uncertainty, constructs a two-stage robust optimization model for configuring energy storage capacity, and calculates the energy storage capacity configuration and daily cost of the virtual power plant under different uncertainty values ​​and prediction error values ​​through the column constraint generation method. Through the present invention, virtual power plant operators can reasonably configure energy storage capacity based on their own acceptance of different prediction confidence levels, thereby improving the rationality and economy of virtual power plant energy storage configuration.

Claims

1. A virtual power plant energy storage configuration method taking into account source and load uncertainty, characterized in that: The steps include: Step 1: Construct a deterministic model for the energy storage capacity configuration of the virtual power plant. The model aims to minimize the comprehensive daily operating cost of the virtual power plant. The constraints include energy storage power station constraints, photovoltaic output constraints, and industrial and commercial user load constraints. Step 2: Based on the deterministic optimization model constructed in step 1, introduce uncertainty factors and convert it into a two-stage robust optimization model with uncertainty factors. The first stage problem of the two-stage model is the outer min problem, and the second stage model is the inner max-min problem. The goal of the model is to find the worst uncertainty scenario and calculate the most economically efficient energy storage capacity configuration plan under the worst scenario; Step 3: Based on the duality principle, the two-stage max-min problem of the two-stage robust optimization model established in step 2 is converted into a dual problem, so that the original two-stage max-min problem becomes a max-max problem and is merged into a single max problem, thereby simplifying the entire model and making it easier to handle and solve; Step 4: Use the column constraint generation algorithm to solve the two-stage distributed robust optimization model to obtain the constructed virtual power plant energy storage configuration result.

2. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 1, characterized in that: In step 1, the objective function of the energy storage capacity configuration model of the virtual power plant is: min(C1+C2-C3+C4-C5+C6)(1) Among them, C1 is the energy storage construction cost; C2 is the energy storage operation and maintenance cost; C3 is the recovery value of energy storage; C4 is the cost of industrial and commercial users of virtual power plants purchasing electricity from energy storage; C5 is the income of self-built photovoltaic power plants of industrial and commercial users of virtual power plants selling electricity to energy storage power stations; C6 is the cost of industrial and commercial users of virtual power plants purchasing electricity from the power grid; The expression of the construction cost of virtual power plant energy storage is: C1=c p P max +c e E max (2) In the formula, c p and c e Respectively represent the unit power cost and unit capacity cost of energy storage equipment; P max Indicates the rated power of energy storage; E max It indicates the rated capacity of energy storage; The operation and maintenance costs of the energy storage system mainly include labor costs, maintenance expenses, and the replacement costs of some components of the energy storage unit. The expression is: Where c0 is the annual operation and maintenance cost per unit charge / discharge power of the energy storage system; Y is the service life of the energy storage system; t r is the inflation rate; d r is the discount rate; The recycling and reuse value of the virtual power plant energy storage device is directly related to its initial purchase cost, and its expression is: C3=γC1 (4) In the formula, γ is the recycling value coefficient; The comprehensive daily cost of virtual power plant energy storage should also include: the cost of industrial and commercial users of the virtual power plant purchasing electricity from energy storage, the income of self-built photovoltaic power plants of industrial and commercial users of the virtual power plant selling electricity to energy storage power stations, and the cost of industrial and commercial users of the virtual power plant purchasing electricity from the power grid. The expressions are: in, is the power supplied by the energy storage to industrial and commercial user i, ω1(t) is the price of electricity purchased by the user from the energy storage; is the power supplied by PV j to the energy storage, ω2(t) is the price of electricity purchased by the virtual power plant energy storage from the user's PV; is the power supplied by the grid to industrial and commercial user i, and ω3(t) is the price of electricity purchased by the user from the grid.

3. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 2, characterized in that: In step 1, the constraints of the energy storage capacity configuration model of the virtual power plant are as follows: The constraints related to virtual power plant energy storage are as follows: 0≤P ch (t)≤U ch (t)P max (8) 0≤P dis (t)≤U dis (t)P max (9) U ch (t)+U dis (t)≤1 (10) P max =0.5E max (13) E(0)=0.4E max (14) Formulas (8)-(9) are the upper and lower limits of the charging and discharging power of energy storage, where P ch (t), P dis (t) is the energy storage charging and discharging power, P max is the maximum charge and discharge power of energy storage; Formula (10) is the charge and discharge state constraint, U ch (t), U dis (t) is a Boolean variable, which indicates the energy storage charging and discharging state, and is used to ensure that the energy storage cannot be charged and discharged at the same time. The variable is 1, which indicates that the energy storage is charging or discharging, and vice versa. Formula (11) is the capacity conservation constraint of the scheduling period, η is the energy storage charging and discharging efficiency, N T is the scheduling period; Formula (12) is the remaining capacity constraint of energy storage in different time periods, E(0) is the initial capacity of the energy storage power station, E max is the maximum capacity of the energy storage station; Formula (13) is the energy storage capacity and power configuration constraints; Formula (14) indicates that the daily initial power of the energy storage station is 0.4 times its rated capacity; Formulas (15)-(16) are the balance constraints of the charging and discharging power of the energy storage station, The power that the grid supplies to the energy storage power station, is the power supplied by photovoltaic power to energy storage, The power supplied by the energy storage power station to industrial user i; The photovoltaic output constraints are as follows: Formula (17) is the power balance constraint for each commercial and industrial user's photovoltaic output; Formulas (18)-(19) are the upper and lower limit constraints of each power, where P m,max is the upper limit of the tie line power; The load constraints for industrial and commercial users are as follows: Formula (20) is the power balance constraint of each industrial user load, that is, the user is powered by its own photovoltaic power, purchases electricity from energy storage and purchases electricity from the grid; Formulas (21)-(22) are the upper and lower limit constraints of each power.

4. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 3, characterized in that: In step 2, the two-stage robust optimization model containing uncertain factors is as follows: Under the premise that the uncertainty factors of photovoltaic output and load power are not considered, a deterministic optimization model for energy storage configuration problem is constructed, and its compact form is as follows: In the formula, x and y are optimization variables, and the specific expression is shown in formula (24); After rewriting equation (1), we get the coefficient column vector form of the objective function, which is c; A, B, C, D, E, and F are the corresponding coefficient matrices under the equality and inequality constraints respectively; a, b, c, d, e, and f are the constant column vectors corresponding to the constraints respectively, and y are uncertain variables, the details are as follows: Where P pvA (t),P pvB (t),P pvC (t),P pv1 (t),P pv2 (t) represents the predicted value of photovoltaic output at time t, and P pv,j (t) Replace; P user1 (t),P user2 (t) represents the predicted value of the charging station load power at time t, and P user,Y (t) Replacement.

5. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 4, characterized in that: In step 2, the two-stage robust optimization model containing uncertain factors is as follows: In the formula, the first-stage problem is the outer min problem, and the optimized variable is x; the second-stage problem is the inner max-min problem, and the optimized variables are u and y; the minimization problem is to reduce the user's daily cost; the goal of this model is to find the most economically efficient energy storage capacity configuration when u takes the worst case value u∈U, that is, to calculate the minimum cost under the worst case; Formula (27) is the uncertainty factor introduced in the original model. By defining the box-type uncertainty set U, the fluctuation range of photovoltaic power generation output and charging station load power is defined, P pv,j (t) and P user,Y (t) is the uncertain variable of photovoltaic output and charging station load introduced after considering the uncertainty, and Represents the fluctuation range of the corresponding prediction value error; Formula (28) indicates that when a set of (x,u) is given, the specific value interval of the optimization variable y is Ω(x,u).

6. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 5, characterized in that: In step 3, the decoupled main problem of the two-stage robust optimization problem is as follows: Where k is the current number of iterations; y l is the solution of the subproblem after the lth iteration; u l * It is the value of the uncertain variable u in the worst scenario found after the lth iteration.

7. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 6, characterized in that: In step 3, the decoupled sub-problems of the two-stage robust optimization problem and its dual problem are as follows: In the formula, Ω(x,u) is the value range of the decision variable y, γ, λ, ν, π, β represent the dual variables corresponding to each constraint in the subproblem, and the max-min problem in the subproblem is converted into an equivalent max problem based on the duality principle, thereby simplifying the structure of the model. The subproblem after the conversion and the dual model are as follows: The uncertain variable corresponding to the optimal solution of the dual problem is a point in the uncertainty set U. When equation (32) takes the maximum value, the value of the uncertain variable u should be at the boundary of the interval, that is, the worst scenario. Therefore, the defined uncertainty set is transformed into the following: In the formula, B=[B pv,j (t),B user,Y (t)] T is a binary variable. When the value is 1, the uncertain variable takes the boundary value of the corresponding period in the interval; Γ pv,j With Γ user,Y It is an uncertainty value, indicating the cumulative number of photovoltaic output and charging station load reaching the boundary of the fluctuation range, and its value range is 0~N T All integers within ; Bringing the transformed uncertainty set into the transformed subproblem, we get the complete dual subproblem, as follows: In the formula, Indicates the fluctuation range of an uncertain quantity.

8. The method for configuring energy storage in a virtual power plant taking into account source and load uncertainty according to claim 7, characterized in that: In step 4, the model is solved using the column constraint generation algorithm as follows: Based on the convergence principle of the column constraint generation method, the main problem and subproblems are solved by calling the GUROBI loop iteration on the MATLAB platform. When the upper bound and the lower bound of the original problem meet the convergence principle at the same time, the solution is completed and the energy storage configuration capacity E is obtained. max , P max , the solution process is: (1) Set LB = -∞, UB = +∞, number of iterations k = 1, and convergence accuracy ε ≤ 0.01; (2) Solve the main problem and find the optimal solution Update LB = α; (3) Given a solution Get the objective function value of the subproblem and the value of uncertainty u in the corresponding worst scenario Update upper bound (4) When |UB-LB|≤ε, terminate the iteration and obtain the optimal solution; otherwise, set k=k+1, add the relevant constraints to the main problem, and repeat steps (2) to (4); Wherein, the termination condition is |UB-LB|≤ε.

Images